Markers at birth may help predict type 1 diabetes

One day, there could be a new test to screen for type 1 diabetes, now that scientists have found markers in the blood of the umbilical cords of children who were later diagnosed with the disease.

In new research published in the journal Nature Communications, scientists at the University of Florida and Linköping University in Sweden found that early signs of type 1 diabetes can be found in the blood of the umbilical cord, which links the baby to the placenta during pregnancy and is discarded at birth.

This does not mean that diabetes is predetermined, said Angelica Ahrens, lead author of the study and assistant research scientist in microbiology at the UF Institute of Food and Agricultural Sciences (UF/IFAS).

“It means that biology is being shaped during a period when systems are still highly adaptable,” said Ahrens. 

Type 1 diabetes is caused when the immune system mistakenly attacks the pancreas’ beta cells, which produce insulin, destroying them. This means the body cannot regulate blood sugar properly, creating a lifelong need for treatment and potentially serious health problems. An estimated 8.4 million people worldwide live with type 1 diabetes, a number expected to increase to 17 million over the next 20 years.

Data from the new study suggest that the process that leads to inflammation starts during pregnancy, Ahrens said. 

“These are processes that could then make biological conditions more favorable for an attack on the pancreas,” she said. “It would be useful if markers in the umbilical cord blood could give doctors a heads-up -- as current screening is genetically based or requires antibodies to the self, at which point the disease has already progressed significantly. Knowing how to offset these processes and the inflammation will be very important in our continued studies of these markers, as they could inform interventions or lifestyle changes in early life.”

Researchers are studying 16,683 babies in Sweden, a group known as All Babies in Southeast Sweden (ABIS). The population-birth cohort follows babies born in a particular region of Sweden between 1997 and 1999. For each child, extensive information across time was provided by the family and later the individuals themselves, including periodic biological samples. The ABIS participants are followed to present day. 

Scientists used machine learning to identify which proteins are related to a risk of developing type 1 diabetes, using the cord blood of a subset of these individuals. Then they looked for patterns in the proteins linked to disease risk, using prediction models, Ahrens said. This analysis involves repeatedly testing and refining models across hundreds of proteins to ensure reliable results. 

To synthesize data across the proteins, researchers used UF’s HiPerGator, the country’s fastest university-owned supercomputer. That made their work efficient, reproducible and scalable, Ahrens said. 

“Access to HiperGator helped move this study forward and positions us well for future work in larger and more diverse populations,” she said.

Within the group studied, scientists discovered protein markers that are observable at birth that are strongly associated with the development of future insulin-dependent diabetes. 

Scientists also saw that some of these markers linked to future disease may be influenced by certain polyfluoroalkyl chemicals, the so-called “forever chemicals,” to which the mother was exposed during pregnancy.

Johnny Ludvigsson, senior professor with the Department of Biomedical and Clinical Sciences at Linköping University and a corresponding author on the study, said the study stems from of a fruitful collaboration between scientists in Sweden and the United States.

"Our study shows that a combination of several factors during pregnancy increases the risk of future development of type 1 diabetes,” Ludvigsson said. “This suggests that no single method will prevent the disease, but different early changes of lifestyle and environment may gradually make type 1 diabetes less common.”

Eric Triplett, professor and chair of the UF/IFAS Department of Microbiology and Cell Science and a corresponding author of the study, said: “The primary value of this paper is that it suggests a new means to predict future type 1 diabetes using a biological tissue, the cord blood, that is discarded at birth. This tissue contains blood from the child that can be stored for future measurement of proteins in the sample.”

“This represents an approach that doesn't require an invasive procedure for the newborn child (such as blood sampling), genetic testing, which has privacy concerns, or any future expensive testing in the child for autoantibodies in blood. We see a bright future for cord blood proteomics in disease prediction.”